多通路空间声的前方四扬声器局域Ambisonics信号馈给法*

典型的多通路空间声扬声器布置一般包含水平面左前、右前,高仰角左前上、右前上四个方向的扬声器。 本文提出一种利用该四个扬声器产生前方水平与垂直方向虚拟源的一阶局域Ambisonics 信号馈给法。该信号馈给法是通过对目标和重放声场进行球谐函数展开并取一阶近似得到。采用简化的头部模型和精确的头相关传输函数模型分析表明,一阶局域Ambisonics 信号馈给法可以产生合适的低频听觉定位因素,包括双耳时间差及其随头部转动的动态变化。虚拟源定位实验结果表明,该方法可以在扬声器布置的范围内,甚至在略超出扬声器布置的范围内产生不同方位角和仰角的虚拟源。因而本文的方法可用在多通路空间声重放中产生与图像配合的虚拟源定位效果。

Local Ambisonics signal mixing method for four frontal loudspeakers in multichannel spatial sound

Typical loudspeaker configurations in multichannel spatial sound usually involve loudspeakers in left-front and right-front directions in the horizontal plane as well as in left-front-high and right-front-high directions in a high elevation. Using these four loudspeakers, the present work proposes a method of the first-order local Ambisonics signal mixing to recreate horizontal- and vertical-frontal virtual source in spatial sound reproduction. The signal mixing is derived by decomposing the target and reproducing sound filed with spherical harmonic functions and matching them up to the first order. By using a simplified head model and an exact HRTF model, it is proved that the signal mixing method can create appropriate directional localization cues at low frequency, including interaural time difference and its dynamic variation with head turning. A virtual source localization experiment indicates that the signal mixing can recreate virtual source in various azimuths and elevations within or even a little bit beyond the range of loudspeaker arrangement. Therefore, the method in present work is applicable to recreate virtual source matching with picture in video-audio reproduction.